Their central insight is that a decentralized crowd can bring greater scale, diversity of expertise, and variation in approaches than even a sophisticated internal organization.
More than a decade later, generative artificial intelligence extends their argument in a powerful new direction. The strategic question is no longer simply whether a company should make, buy, or crowdsource innovation.
Managers increasingly face a broader choice:
What should be solved internally, what should be opened to a crowd, what should be delegated to AI, and what should be solved through a combination of all three?
That may become one of the defining organizational design questions of the AI era.
The Crowd Solves the Search Problem
One of the most important advantages of crowdsourcing is not simply cheaper labor. It is the ability to search a much larger solution space.
Boudreau and Lakhani argue that when a problem has a well-defined objective function—meaning you can clearly evaluate whether a proposed solution is better or worse—crowds can be remarkably effective, because each participant explores a different part of the solution landscape independently.
This is what makes contests and open innovation challenges powerful. The winning solution often comes from an unexpected direction, proposed by someone whose expertise lies outside the field where the problem originated.
The authors cite examples from organizations like NASA, Eli Lilly, and Procter & Gamble, where crowdsourcing produced breakthroughs that internal R&D teams had not achieved.
But crowdsourcing has limitations. The authors note that the crowd model works best when the problem can be decomposed into independent modules, when intellectual property considerations are manageable, and when the cost of evaluating submissions is low relative to the value of the solution.
For problems that require deep integration across disciplines, sustained investment over long time horizons, or protection of proprietary knowledge, internal innovation remains superior.
AI Changes the Innovation Landscape Again
Generative AI introduces a fundamentally new participant in the innovation ecosystem.
Unlike a human crowd, AI can generate solutions at enormous speed, iterate on feedback almost instantly, and operate continuously without fatigue. Unlike an internal team, AI does not have institutional memory, political incentives, or professional identity that might limit the range of solutions it considers.
The practical implication is that managers now face a three-way decision rather than a two-way one.
